1. Software art, generative art, the aesthetics of programming 

Over the last few years I have programmed hundreds of computers to produce flows of random images and have left them working endlessly in my installations. What is the relationship between these works and generative art, software art and the other manifestations that can be traced back to the aesthetics of programming?
The production of images through programmed machines covers a range of phenomena which are not always artistic in intent. A partial list could include:
- computer graphics, those produced by artists using commercial software as well as those produced by software artists;
- generative design, which uses programs capable, through randomisation and morphogenetic approaches, of creating alternative models and solutions, i.e. variants to be evaluated and selected (evaluating and selecting are terms that, according to some artists, recall natural evolution [1] and can define a new paradigm in art that is linked to the aesthetics of programming);
- the production of generative content as a contribution by visual artists to the performances of dancers, musicians or (as in the work presented by Golan Levin at the 2003 edition of Ars Electronica) vocal performers;
- the work of VJs who use generative technologies in discotheques in order to create or modify images that in some cases may vary according to the characteristics of sound, providing an example of pop synaesthesia;
- videogames, whose software often uses non-deterministic algorithms, i.e. procedures that can produce many different unforeseeable outcomes starting from the same input.
The aesthetics of programming (I shall consider here only the production of images) regards phenomena which can be evaluated on different levels. All of them, however, share two common elements: 1) the delegation of the process to a machine and 2) the (partial) abandoning of control over outcomes. These two aspects are inextricably linked. In discotheques as in museums, the aesthetics of programming is the aesthetics of your action which can be made independent and indefinitely expanded, and this can happen only on condition that there is a certain loss of control. It is in these terms, I believe, that we can consider both my Image Machines (1988-) and Sealed Computers (1992-), machines which I program to produce flows of random images and which I then leave to work indefinitely [2-6]. 
The most obvious distinction between these works and many manifestations of the aesthetics of programming is a question of the limited importance given to images. My works do not start from images but from the process that generates them. The essential dimension of my works lies in the device (which, however, does not only consist of hardware and software because, in some interactive works, it can also involve the public itself). Delegating to technological devices – for instance, using randomness in an extreme way – means opening up new research possibilities. Above all it means starting from the disproportion between the artist and the processes which are activated by his or her work, which for the first time tend to go beyond the artist himself: on the one hand, there is your action which can be expanded indefinitely; on the other, there is your self-reduction to the role of spectator of uncontrollable processes. The problem of the relationship between infinity and identity, which in generative art is often discussed in terms of the identity (oneness and identification) of individuals (images, for example) belonging to the same species, is translated at a deeper level into the problem of the identity (individuality) of the artist and his or her subjectivity. These are aspects that Mario Costa [7-8] has investigated at great length, reaching the conclusion that “the strong categories of aesthetics in the neo-technological age are no longer ‘interiority’, ‘expression’, ‘meaning’, ‘style’, ‘artistic personality’, the ‘symbolic’ and so on, but ‘exteriority’, ‘signifiers’, the ‘non-subject’ and the ‘physiology of the machine’”. 
I am aware that my work goes against the grain. Most software artists (a definition that puts the emphasis on the instrument while neglecting the intentions of users) continue to work on the quality of images. The situation is similar to that which was earlier experienced in electronic music (which needed less powerful machines): dozens of generative programs, capable of producing algorithmic compositions and modulations of timbre, were initially used experimentally and later in a general way. Now in visual art, too, anybody can acquire software and use it to generate any kind of graphic output. 
This spread of generative technology has consequences which are not only positive. New technologies have brought art to a decisive turning point, but the crucial problem of artistic research still remains the same: we need to distance ourselves from useless things and from exhausted poetics, which tend to reappear constantly because they are more easily sustained by the market and often make use of new technological devices as a way of disguising their anachronistic qualities. We need to avoid letting the spread of software art produce an aesthetic art which is too peaceable and entrenched in the present to be of any use to us, an electronic baroque which we could readily do without.
I don’t think, however, that this can be avoided. For some years technological innovation has been outpacing the art system. Works using new technologies in a more critical spirit have not yet been assimilated, and there is continuing confusion between different kinds of poetics.
Within this framework of fast but predictable development, I have continued to work with my Computer sigillati since 1992, considering them as a purely quantitative installation, absolute and empty and far removed from the rest: hundreds of machines programmed to generate kilometres of images, indefinitely, independently, and in the absence of any public.

2. Delegating to machines: images and superimages, quality and quantity

Perhaps it would be better to call what I produce with my machines not images but superimages, i.e. flows of images that are subject to continuous variations. In some cases these are produces by devices including not only the machines and their software, but also the public, which can interfere in their functioning. In any case, my approach is always quantitative: I want the drawings produced by my machines to cover endless surfaces; their characteristics and quality are of less interest. I never thought of codes or devices in terms of style, but rather as kinds of DNA.
In the initial phase, at the end of 80s, in spite of the limited power of computers at that time, I tried to produce images whose formal quality would simulate the existence of an artist. But I soon abandoned this approach. It seemed to me that there was something so banal about the images (even those produced by a machine) and the formality of their beauty as to put them beyond redemption. I thought that this was due not only to the over-abundance of images produced by video technologies but also to the characteristics of digital technologies themselves, which are capable of activating processes beyond the representation of reality and of acting directly on the way it functions, thus relegating the linguistic mediation of image to a secondary role.
Sometimes I program these machines to produce images that have certain characteristics, but working on the quality of images interested me only as a way of studying certain preferences. For instance, the fact that we find images more attractive according to the index of their structural complexity is interesting. Highly ordered and disordered systems fail to work well, because both, for opposite reasons, possess limited structural complexity. This point needs to be emphasised because it is one that has been examined with particular interest by students of generative art [9]. Not always does the “algorithmic complexity” of a system correspond to its “effective complexity”. A system generating merely random outcomes (images) may have high algorithmic complexity but low effective complexity, which is limited in both highly ordered and disordered systems, while it is higher in intermediate systems since it is a function of two different things, structure and random variation. Sometime I take this into account when programming my Image Machines where images are visible. But even when I think of the parallel (not visible) information universe of signs drawn by my sealed machines, it pleases me more to know that a complex and intricate landscape is being created.
In this perspective we can also consider some works based on the application of artificial intelligence. AIMS (Artificial Intelligence Mediated Sublime, 2003) uses a genetic algorithm, conceiving images as a "territory" inhabited by various populations of signs (also in this case white lines on a black background), in competition with each other and capable of reproducing themselves and their DNA.
The question of the quality of images is also implied by works such as SMSMS (SMS Mediated Sublime, 2002), from the CIMs (Collective Intelligence Machines) series [8, 10-12], and Atlas 2 (2001). In the first, the public can give a direction to the process generating the images and can change some of their characteristics. In this case too, however, the centre of the work is not the image but the device, expanded to include participants and the mobile telephone network they use to interfere with the otherwise unforeseeable working of the machine. The centrality of the device is even more evident in the most recent versions of this series, where I use the telephone network also in order to exercise remote control and reshape the process (CIM 2), as well as to connect multiple, geographically distant, installations (CIM 3).
In the Atlas 2 series, I allow non-artist programmers to modify a piece of software previously used to program some machines. The code is passed on without any control: participation is open and the only condition is that the endless continuity of the flow of images be maintained. There are series, started in Moscow and in Bangalore, which have grown rapidly, in some cases with minimal variations, in others with inventions that introduce radical changes.
In my other works I have focused on purely quantitative production: I use my machines to produce images which are limitless in time and space. Also my Computer sigillati, programmed and working machines whose images are not visible, are not simply conceptual works. Since the very early stages of this work I have wanted these machines to be capable of a certain speed of execution (computers at the time were slow and expensive so, since I needed large numbers of them, I was always looking for old computers which already had a mathematical co-processor installed). I was not then, and am not now, interested in being considered an artist who creates certain images, nor do I want to be simply a conceptual artist. Rather my aim is to be an artist who, through his machines, actually traces more lines than anyone else, covering endless surfaces.

3. Giving up control: randomness, collective intelligence, artificial intelligence

Giving up control, delegating one’s choices to certain procedural rules are ideas that have a long history in artistic research. They can be traced back not only to the standard references of Dadaism or John Cage, but also to Conceptualism (for example, Sol Le Witt’s combinations of basic forms determined by external rules, or Alighiero Boetti’s permutations). A programmed machine, though, goes beyond these procedures, however autonomous they may be. The point is not only to lose control and delegate choices to some procedure or other (Tzara’s random extraction of words, or the dropping of a thread in Duchamp’s Large Glass). There is a leap in quality that is due solely to technology, which for the first time enables us to become spectators of these processes, creating works that previously could only be hypotheses, or metaphors, and shifting the barycentre from the artist who imagines things to the devices that create them autonomously. This is why I leave my machines working indefinitely. And when their outcomes are not purely quantitative, it is because I have delegated some choices to the public – as in the CIMs, through the use of techniques of collective intelligence – or I have put choices in the hands of some programmers – as in Atlas 2 – limiting my own role to organizing the operation.

I would like to clarify this aspect by pointing out the ways in which results that are out of my control (in most cases images) are generated in my works. Delegating this process to a device is possible by adopting two different approaches: 
1) the use of algorithms capable of making random choices (randomisation): any computer can generate pseudo-random numbers starting from a given numerical series which can be activated from various points each defined by a random event (for example, time measured in milliseconds);
2) the introduction of an evolutionary principle which transfers intelligence to the system; this can be done in two further ways: through the application of artificial intelligence or collective intelligence. In the former case, programming techniques (genetic algorithms, neural networks etc.) are used to develop different possible solutions according to their fitness to given objectives. In the latter case, procedures are applied which enable the public to interact and become part of the device.
My IMs and the Computer sigillati use randomisation, generating a merely statistical, directionless randomness, either by using the pseudo-random numerical series contained in any computer or, in some cases, taking the necessary random values from outside, as in the NAA (Nikkey Assisted Art, 1997) series, a work in which I used fluctuations in the Tokyo stock exchange index. 
Another interesting direction in artistic research is the possibility to apply collective intelligence and group communication techniques. Collective Intelligence Machines are interactive installations where some of my computers, programmed to produce random images, are connected to the mobile telephone network. This allows the public to interact with the machines by sending SMS messages that update the parameters of the system and therefore the characteristics of the images. Through a process of asynchronous communication (based on repetition and feedback, similar to some techniques used in electronic democracy applications [13]), the installation offers the public an interface across which the images, which are projected onto large walls, can react to their continuous input.

After reviewing these different works, the point that I would like to underline in conclusion is that new technologies imply a change in the content of artistic research. It is true that these technologies can be used (forced to realize different intentions) similarly to others, but what would be the purpose (to go back to the initial point) of, for example, continuing to put the emphasis on images rather than the devices generating them (or – to put it differently – of continuing to consider images as static instead of dynamic systems, indeed as devices themselves)? These questions cannot be ignored (it would be as if an action painter, after making his action autonomous and multiplying himself thousands times, had settled for exhibiting coloured surfaces). The point to underline is that with the new technologies artistic research is finally moving from the representation of reality to the functioning of reality. This process is also determining a change in the role of artist, who is being transformed into a kind of “actor of complexity”: representation is by definition something that is appropriate to the artist, but today, with the new technologies, precisely what is not appropriate seems capable of becoming art, transforming the very disproportion between the artist and the processes activated by his or her work into the content of a new area of experimentation.

References

1. C. Soddu, "Visionary Aesthetics and Architecture Variations", Proceedings GA2003.
2. D. Scudero (ed.), Maurizio Bolognini: installazioni, disegni, azioni (on/off line), Lithos, Roma 2003.
3. M. Bolognini, Atlas, Nuovi Strumenti, 2001.
4. M. Costa, "Maurizio Bolognini and technological randomness", in Macchine casuali, Nuovi Strumenti, 1997.
5. E. Pedrini "New Technologies and Linguistic Evolutions in Contemporary Art", in Approaches in Multimedia Art, Lubelski, New York 2003.
6. D. Scudero, “Arte e tecnologie digitali in una conversazione con Maurizio Bolognini”, Luxflux, no. 6, 2004.
7. M. Costa, Il sublime tecnologico, Castelvecchi, Roma 1999 (1st ed. 1990).
8. M. Costa, New Technologies, Artmedia, Museo del Sannio, Benevento 2003.
9. P. Galanter, "What is Generative Art? Complexity Theory as to Context for Art Theory", Proceedings GA2003.
10. M. Bolognini, “The SMSMS Project: Collective Intelligence Machines in the Digital City”, Leonardo, MIT Press, Vol. 37, No. 2, 2004.
11. M. Bolognini, “SMSMS (Short Message Service Mediated Sublime): Infoinstallations et ville numerique”, Ligeia, No. 45-48, 2003.
12. R. C. Morgan, “Maurizio Bolognini: la problematica dell’arte”, Luxflux, No. 6, 2004.
13. M. Bolognini, Democrazia elettronica, Carocci, Roma 2001.

 |intro  index|